Fast Discovery of Reliable Subnetworks

Abstract

We present a novel and efficient algorithm, Path Covering, for solving the most reliable subgraph problem. A reliable subgraph gives a concise summary of the connectivity between two given individuals in a social network. Formally, the given network is seen as a Bernoulli random graph Q, and the objective is to find a subgraph H ⊂ Q with at most B edges such that the probability that a path exists in H between the given two individuals is maximized. The algorithm is based on an efficient stochastic search of candidate paths, and the use of Monte-Carlo simulation to cast the problem as a set cover problem. Experimental evaluation on real graphs derived from DBLP bibliography database indicates superior performance of the proposed algorithm.